WO2007110107A1

WO2007110107A1 - Method for the three-dimensional measurement of fast-moving objects

Info

Publication number: WO2007110107A1
Application number: PCT/EP2006/061017
Authority: WO
Inventors: Bernward Mähner
Original assignee: Maehner Bernward
Priority date: 2006-03-23
Filing date: 2006-03-23
Publication date: 2007-10-04
Also published as: ATE508342T1; US20090244261A1; EP1996897B1; EP1996897B2; EP1996897A1; DE502006009454D1

Abstract

Light sectioning or fringe projection methods are used to measure the surface of objects (2). According to said methods, the object (2) is moved past a measuring system (4) and the measured data is recorded during this movement. These methods can only be used for a high-resolution, comprehensive measurement of the surface if the object (2) is moved sufficiently slowly, in relation to the scanning rate of the measuring system (4), past said measuring system (4). To achieve a comprehensive measurement of the object surface, even when the object (2) performs a relatively fast movement, in particular a rotational movement, the object (2) is repeatedly moved past the measuring system (4) and measured. The use of a trigger device for recording the measured values permits the data obtained in the individual passes to be correlated in a three-dimensionally correct manner in relation to one another and the surface of the object to be measured with high resolution.

Description

Method for the spatial measurement of fast-moving objects

It is known that for the spatial measurement of objects triangulation systems consisting of a point laser and a line scan camera are used. Such systems are used for example in lathes to continuously check certain dimensions or tolerances in the production of rotating bodies. Due to the principle, only one point is detected on the object at a time, but several systems can be used in parallel. The line rate of commercially available line scan cameras is up to over 200 kHz, so that the distance between the individual measuring points is very small, even in the case of rapidly rotating workpieces. Disadvantage of the method is that even when using multiple systems only a relatively small number of measuring points is detected simultaneously.

This disadvantage is overcome in the light section method. By using a line laser instead of the point laser and a surface camera instead of the line camera, an additional dimension is spanned, so that instead of a measuring point, an entire contour line is detected. In contrast to the line scan cameras, the frame rate of commercially available area cameras with 25 to 60 frames per second is comparatively low. The object to be measured must therefore be moved sufficiently slowly past the light-section system in order not to exceed a usable spatial distance between the individual measurements. Such a system is described for example in the patent DE 100 19 386 C2. In the embodiment shown here, the object rotates relatively slowly with a frequency of about 0.5 Hz about the axis of rotation. In contrast, the frame rate of the video camera used is about 25 Hz and is therefore sufficiently high. However, in order to achieve a comparable frequency ratio when measuring fast-moving objects, for example a tire rotating at a frequency of about 10 Hz on a chassis dynamometer, a significantly faster camera with a frame rate of about 500 Hz must be used. For such high-speed cameras, however, considerable additional costs have to be accepted in comparison to standard cameras.

In the fringe projection method, instead of a line laser, a fringe projector is used, so that instead of a contour line, an entire surface can be topometrically detected. For this purpose, a whole series of different projection methods is known. For example, the patent DE 38 43 396 C1 describes a method in which only a single stripe pattern has to be projected for the measurement.

Both laser triangulation and fringe projection methods are subject to the shading problem in principle, that is to say points on the test object that are detected by the camera, but can not be measured due to the object's tomography not simultaneously illuminated by the laser or projection beam. In order to minimize this problem, it is known from the published patent application DE 197 41 730 A1 to pass the object with a rotating or longitudinally oscillating movement several times on the sensor and at the same time to let it execute a longitudinal oscillation or a rotational movement. As a result, the individual surface areas of the object for the measuring system can be detected from a plurality of different observation directions and the shading problem is thereby alleviated. From technical applications, a stroboscopic illumination is also known for the observation of fast-moving or oscillating objects. The flash illumination freezes the image of the observed object for the observer to a specific phase within a movement period of the object. This is also used for metrological purposes. To check a correct ignition setting on a gasoline engine, for example, a stroboscopic lamp is triggered by the primary circuit of the ignition. If in this way the engine flywheel is illuminated with the position marks of the crankshaft, then it is possible for the observer to check the ignition timing with respect to the crankshaft position with the engine running. Another use case for the stroboscope is the analysis of oscillating objects. Here, the phase position of the stroboscopic flash is continuously shifted with respect to the periodic object movement, whereby the oscillatory shape of the object can be observed in quasi-slow motion, since now the oscillation of the object becomes visible to the observer with the speed of the phase position change.

In order to be able to analyze the deformation of an object with periodic, in particular sinusoidal, excitation with a fringe projection method, DE 198 41 365 C2 proposes exposing the camera of the measuring system in a fixed phase with respect to the excitation of the oscillation and with a stroboscopic short Exposure time trigger. In this way, a phase-dependent deformation state of the dynamically deforming object is "frozen" and made available for quantitative evaluation, but to capture a quasistatic state of the object, the exposure time must as a rule be chosen so short that the camera image is underexposed The exposure of the camera is then triggered in several successive oscillations with the same phase angle, thereby accumulating the light energy incident on the sensor of the camera and resulting in a well-controlled image. Based on this prior art, it is an object of the invention to provide a method which allows the measurement of fast-moving objects by means of light-section or fringe projection method. The process should thereby avoid the use of special equipment and the associated high costs and rather be limited to the use of commercially available components.

This object is achieved by a method having the features according to claim 1. Preferred embodiments of the method according to the invention are defined in claims 2 to 18.

According to the invention, a 3D measuring device is used which comprises a projection unit for one or more light-section planes and a surface camera. The projection unit preferably has a line laser or a projector for projection of stripe patterns. The area camera is preferably an electronic CCD or CMOS camera.

According to the invention, the object is moved past the sensor several times, that is, in several passes. Such a movement can take place, for example, by a rotation of the object about an axis of rotation which is stationary relative to the measuring device. During a passage, according to the invention, only individual partial areas of the object surface are respectively detected by the sensor, that is to say a smaller number of partial areas in relation to the total number of all partial areas that can be detected during a complete measurement. Several passages are now carried out according to the invention in such a way that the partial areas of the object surface detected in the individual passages are spatially offset relative to one another with respect to the object. For a new run, the object positions for the measurement data acquisition are selected in advance so that subranges of the Object surface are recorded, which have not (yet) been detected in the previous rounds. This procedure is applied for as many passes until the surface of the object has been acquired at the desired sampling density. The detection of the object surface is thus distributed over a total of several passes, so that even with a relatively slow-acting measuring system, the surface of the object can be detected with a high sampling density.

If the spatial assignment of the contour data obtained from the acquired subregions to one another is not required, that is to say no surface model of the object is required, then the contour data obtained need not be transformed into a common object coordinate system. If the object carries out a periodic turning or oscillating movement relative to the measuring device, in this case the method according to the invention is advantageously used in such a way that the camera is operated with the highest possible image frequency or image exposure frequency. The image or image exposure frequency is so detuned from the movement frequency of the object that the object positions detected in the individual camera images, over which movement periods of the object are considered, travel along the object. After a sufficient number of movement periods of the object, this results in a nationwide detection of the object contour without further technical measures.

An application example of this embodiment of the invention is, for example, the measurement of rims or tires in which the concentricity properties are to be tested, so only maximum height and side impact are of interest. In this application, it is sufficient to calculate, for example, the extreme values determined from all acquired spatial coordinates in the axial and radial directions. In a further optional method step, on the other hand, the contour data obtained from the acquired subareas are transformed into a common object coordinate system and from this a surface model of the object is created. This process step requires precise knowledge of the position of all detected sub-areas with each other and with respect to the object. This information is obtained according to the invention in that the respective instantaneous position of the object is registered for each acquisition of measurement data by the camera of the measuring device.

The determination of the instantaneous positions advantageously takes place via a position transmitter. In an embodiment of the method which is suitable for all rotary or oscillatory movements of the object, it may be a rotary encoder or, in the case of translational movements, a linear displacement measuring system, for example during rotary movements. In both cases, the exposure of the camera is then triggered in each case after reaching a certain value determined by the position sensor and the determined value is stored for further processing together with the image data.

In the presence of a periodic rotary or oscillatory motion of the object, the method is advantageously carried out in such a way that the individual measuring positions are selectively selected via a time control. For the exact timing advantageously a real-time calculator, for example, a computer-controlled counter card is used, which triggers the image exposure of the camera at previously calculated times. This embodiment has the advantage that no high-resolution rotary angle or linear displacement measuring systems must be used. If the measuring system has no direct control over the object movement, advantageously a signal generator is used which generates a reference signal for the synchronization of the time control with the object movement. The too A spatial position of the object present at a specific exposure time is then determined from the time interval between receipt of a synchronization pulse in the reference signal and the exposure time.

In order to prevent drifting of the object positions determined by the exposure times from the actual object positions, the time control is advantageously synchronized with the object movement, for example once per period. For this purpose, a signal generator that sends a synchronization pulse per period at a given instantaneous position of the object. This synchronization pulse is then advantageously used to measure on the one hand the currently present period length of the object movement and, on the other hand, to always relate the exposure times to the respective last preceding or first synchronization pulse obtained after the relevant image exposure.

In all embodiments discussed the exact exposure of the camera of the measuring system is essential. In order to expose the camera to the time points calculated, for example, by the image processing computer, according to a further embodiment of the invention, a mechanical or electronic camera shutter which can be triggered via an external signal is used. But it can also be used a camera whose image input can be triggered externally. In a further embodiment, by contrast, the light source of the measuring system, that is to say, for example, a line laser module, is stroboscopically actuated by a mechanical shutter or an electronic switching device. The time interval between two successive exposures for detecting the subregions of the object surface is advantageously always chosen to be greater than the image period of the camera, so that a double exposure of camera images or field images is excluded. hereby It is ensured that there will be no difficulties in the subsequent image analysis required for light-section and fringe projection systems. Since the object is detected in motion, the exposure time is chosen to be sufficiently short and is only a fraction of the frame period for electronic cameras. It is advantageous if the shutter speed is controlled by the camera itself, even if the camera shutter or the image feed is triggered externally.

In the presence of a periodic rotary or oscillatory motion of the object, a further embodiment of the invention is to run the camera freely or to operate at a fixed frame rate and to provide no external, asynchronous triggering of the camera shutter. The frame rate is selected such that the measurement positions, along the movement periods of the object, travel along the object. For this purpose, it is sufficient to ensure that the frame rate and the frequency of movement are in a relationship with each other, which ensures that identical subareas do not repeat until a sufficient number of passes has been reached. Such a relationship is always not complete.

For each camera image, the time interval between the image acquisition and the time at which the object has reached a certain reference position is measured according to the invention. For this purpose, in turn, a reference signal is advantageously generated which generates a synchronization pulse at a specific instantaneous position of the object at least once per period. By measuring the time elapsed between the image acquisition and the reaching of the reference position, a time value is obtained for each image, which in the presence of a periodic object movement is equivalent to the object position. By means of the measured time values, the contour data of the subareas of the object surface detected in the individual images are then converted into a common object coordinate. System transformed. If the exact exposure time of the camera can not be detected, since, for example, the internal shutter function of the camera can not be tapped externally, the time interval can be used, for example, as the time interval between the synchronization pulse of the object movement and the vertical synchronization pulse in the video signal for the relevant camera image , Overall, this leads to a constant offset for all time intervals determined. For measuring the time intervals, it is advantageously possible to use a real-time-capable calculating unit, for example in the form of a counter card, which is integrated, for example, in the computer system of image processing.

In the following, embodiments of the invention will be explained with reference to drawings. In the drawings show:

Fig. 1 shows a system for carrying out the inventive

Method in a first embodiment;

FIG. 2 is a side view of the system shown in FIG. 1; FIG.

Figs. 3, 4, 5, 6 are exposure timing diagrams for the system shown in Fig. 1; Fig. 7 is an exposure timing diagram for a system for

Implementation of the method according to the invention in a second embodiment;

FIG. 8 shows an exposure timing diagram for a system as in FIG

FIG. 7, however, with a different calculation of the exposure times; FIG.

9 shows the position of light sections on an object according to the methods shown in FIGS. 7 and 8 and

10 shows the determination of a shutdown criterion and a process visualization for the method shown in FIGS. 7 and 8. Figs. 1 and 2 show the schematic structure of a test system for carrying out the method according to the invention in the front and side view. On a chassis dynamometer for vehicle tires, a wheel 1 is pressed against a drive roller 3. The drive roller 3 is driven by an electric motor with adjustable speed. The wheel 1 with the tire 2 to be tested is in turn driven by the drive roller 3. Contact force and drive speed are adjustable and thus make it possible to simulate different driving and load conditions. On one side of the tire 2, a light-section system 4 is mounted to measure a sidewall of the tire 2 during the test. The light editing system 4 includes a camera 6 and a line laser module 5. The light editing system 4 is connected to a computer system 7 equipped with image processing for processing the image data supplied from the camera 6. The camera 6 of the light-section system 4 is equipped with an asynchronously actuated electronic shutter, which can be opened by the computer system 7 by means of a control signal sent via a signal line 8. On the axis of rotation 12 of the wheel 1, a pulse generator 9 is mounted, which triggers an impulse when passing a sensor 10. One wheel is triggered per wheel rotation. The pulses of the sensor 10 are detected by the computer system 7. In practical application, another light-section system 4 would be mounted for the simultaneous measurement of the second side wall of the tire 2.

First, the chassis dynamometer is adjusted to a constant speed. On the basis of the pulses sent by the sensor 10, the periodic length of a wheel rotation is determined by the computer system 7 and from this a suitable series of exposure times is calculated. Thereafter, a series of images is taken by the camera 6 at the calculated exposure times, the exposure times being controlled by the computer system 7. For this purpose, the computer system 7 is equipped with a real-time capable counter card which triggers the closure opening on the camera 6 after the passage of programmable time intervals via the signal line 8. JE the time interval is chosen so that the time interval between two successive exposures is greater than the frame period of the camera and thus a double exposure of camera images is excluded. The exposure time is controlled by the camera 6 itself and is so short that the images of the tire 2 are sufficiently sharp. The recording of the image series is terminated as soon as a sufficiently large number of light sections has been recorded, which are distributed as evenly as possible over the circumference of the wheel.

With regard to the maximum number of light sections that can be detected per wheel revolution, there are three cases:

1st case: Low speed, that is per wheel rotation several shots on the circumference possible. 2nd case: Mean speed, that is per wheel turn is one

Recording on the circumference possible.

3rd case: High speed, that is per wheel rotation is less than a recording on the circumference possible.

FIGS. 3 to 7 show the signals or states decisive for the selection of the exposure times. In detail, the curves "Radimpuls", "VD" and "Bildbelichtung" have the following meaning: The curve "Radimpuls" shows the synchronization pulse which is sent once per wheel rotation. The curve "VD" shows the image clock of the free-running, that is continuously recording camera 6. The distance between two pulses of the VD curve corresponds to the image period F. Depending on the type of camera, an image clock corresponds to either a full field or a field. Image exposure "shows at what times an image exposure is triggered on the camera. In the case shown, the image exposure is triggered in the positive signal edge. Fig. 3 illustrates an exposure timing chart for the above-described system for the 1st case, that is, when multiple exposures per wheel rotation are possible. This is the case if the period length T of a wheel rotation is greater than the time between two video images. For the exposure times, two variants are shown. The first (= upper) variant uses exposure times, which are offset by a time interval of t1 = T / 3 within a wheel rotation by 120 °. During the transition from one wheel rotation to the next, an additional time interval dt is waited for in order to ensure that the three light sections, which are recorded within the following wheel rotation, are offset from those of the preceding wheel rotation at the wheel circumference. For example, the time interval dt may correspond to a rotation angle of 0.5 ° with dt = T / 720. After 240 revolutions (= 12070.5 °) the wheel is then detected with 720 light cuts at a distance of 0.5 °.

An alternative determination of the exposure times is shown in the underlying image exposure curve. Constant time intervals t2 between two exposures are always used. For example, the time interval t2 is selected to be equal to (T + dt) / 3. If dt again corresponds to a wheel angle of 0.5 °, the wheel will ultimately be detected in 720 revolutions (= 120 ° / (0.573)) with 2160 light cuts at a distance of 0.573.

Since in the system shown in Fig. 1 between the drive roller 3 and wheel 1 is not a fixed gear ratio, the rotational speed of the wheel 1 may be subject to undesirable fluctuations and slippage. Therefore, it is advantageous to refer the time for the first exposure within a period to the last received wheel pulse. The time interval for the respective first exposure within one period of the wheel rotation then results in nx dt where n is the number of wheel rotations. As a result, the image acquisition is synchronized once per revolution with the wheel rotation. Fig. 4 shows an exposure timing diagram for the system of FIG. 1 for the second case, that is, when the rotational frequency of the wheel 1 is only slightly smaller than the frame rate of the camera 6. The exposure timing diagram shows that only one recording per wheel rotation takes place. The time interval t2 between two exposures is chosen so that it is greater by dt than the period T.

The exposure timing chart shown in Fig. 5 shows the conditions when using the method for the 3rd case, that is, when the rotational frequency of the wheel 1 is larger than the frame rate of the camera 6. In this case, an image can not be taken within each wheel rotation , In the embodiment shown, exposure takes place only in every second wheel rotation. The distance between two exposures results in t2 = 2 × T + dt, where dt is again chosen from the point of view that a sufficiently good sampling rate results after the end of the measurement.

FIG. 6 shows the same relationships with respect to the rotational frequency of the wheel 1 and the image frequency of the camera 6 as in FIG. 5. However, the distance between two exposures has been shortened compared with FIG. 5 to t2 = 1.75 × T + dt. On the one hand, this time interval is still somewhat larger than the image period F between two camera images, on the other hand the measuring time is shortened by 12.5% compared with the embodiment from FIG. 5.

FIG. 7 shows the exposure timing diagram for a system modified from the system of FIG. 1 such that there is no external, asynchronous opening of the shutter of the camera 6. Furthermore, the real-time calculator is used to measure the time intervals between image exposure and wheel pulse. The camera 6 now works with camera-controlled image exposure. Since the object, that is to say the tire 2, is in rapid motion when the method according to the invention is used, the exposure time is chosen to be significantly shorter than the image period F. Such an adjustment of the exposure time is possible in virtually every electronic camera 6 by means of an internal electronic shutter. The shutter is triggered by the free-running camera 6 itself each time after a certain time ts within a frame period F. From the computer system 7, the times t1 to t9 are measured, in each case the time is measured, which has elapsed after receipt of the last wheel pulse and the opening of the shutter by the camera 6. The times t1 to t9 are proportional to the rotational position of the wheel 1 and thus make it possible to arrange the contour lines recorded in the individual camera recordings with respect to the tire 2 in the correct position relative to one another.

FIG. 8 shows the exposure timing diagram for a system which operates analogously to that of FIG. 7, wherein, however, the calculation of the exposure times is modified such that a particularly simple and cost-effective technical realization is possible. Within one revolution of the wheel 1, only the time t1, t4, t8 is measured from the wheel pulse to the first image exposure, for example with a counter card. The times of the image exposures following the wheel rotation are then respectively calculated by adding the number of image clocks between the first and the current image multiplied by the constant image period F and the time value of the first image.

FIG. 9 shows which rotational positions of the wheel from FIG. 2 correspond to the times t1 to t9 from FIG. 7 and FIG. 8, respectively. Furthermore, the drawn rotational position 0 corresponds to the wheel pulse 11 shown in FIGS. 7 and 8.

In a procedure according to the method shown in FIGS. 7 and 8, the position of the individual measurements on the wheel circumference is not necessarily predetermined and the rotational angular distance between two adjacent measurements is not constant. Fig. 10 illustrates how in these cases the measurement is carried out in a controlled manner. In the current measurement are the determined image exposure times t1 to t9 corresponding rotation angle φ _t i to φ _t g continuously sorted by size and then the angular distance dφ between each adjacent rotation angles φ _t calculated. As the measuring duration progresses, these angular distances dφ become smaller and smaller as continuous additional individual measurements and thus intermediate positions are added. The measurement is advantageously continued until the maximum available angular distance dφ _ma χ falls below a predetermined threshold. To visualize the measuring process, the detected rotational positions of the wheel 1 are advantageously entered in a graduated scale 11 and displayed. However, this type of process visualization can also be used in all other embodiments of the invention.

LIST OF REFERENCE NUMBERS

1 wheel

2 tires

3 drive shaft

4 light section system

5 line laser module

6 camera

7 computer system

8 signal line

9 pulse generator

10 sensor

11 degree scale

12 axis of rotation

F picture period

I wheel impulse

T period length

t time interval, exposure time dt time interval n number of wheel rotations ts tripping time

φ rotation angle dφ angular distance

Claims

claims

1. A method for the spatial measurement of fast moving objects (2), wherein the surface contour of an object (2) is determined by the object (2) is measured three-dimensionally by a light-section or stripe projection method, characterized in that the object (2 ) in several

Passages on a measuring device (4) is moved past, wherein of the measuring device (4) in the individual passages each subregions of the object surface, which are spatially offset with respect to the objects detected in the other passages with respect to the object (2), and this offset being generated by the object surface being picked up by the measuring device (4) at different instantaneous positions of the object (2).

2. Method according to claim 1, characterized in that the instantaneous position of the object (2) present in each case during the detection of the subregions of the object surface is registered with respect to the measuring device (4), the contour data of the detected subregions being integrated into a common object by means of the registered instantaneous positions. jektkoordinatensystem be transformed and from the transformed contour data, a surface model of the object (2) is generated.

3. The method according to claim 2, characterized in that the instantaneous positions of the object (2) relative to the measuring device (4) in each case by a measurement of the object (2) relative to the

Measuring device (4) traveled distance or of the object (2) relative to the measuring device (4) covered angle of rotation (φ) are determined.

4. The method according to any one of claims 1 to 3, characterized in that the object (2) relative to the measuring device (4) carries out a periodic rotational movement or a periodic oscillatory movement.

5. The method according to claim 4, characterized in that a reference signal is generated which at least once per period a synchronization pulse (h to li ₆ ) sends, preferably by the transmitted synchronization pulses (I ₁ to I ₁₆ ) the image exposure of a camera (6 ) of the measuring device (4) is synchronized with the object movement.

6. The method according to claim 5, characterized in that the currently present period length (T) of the rotary or oscillating motion is determined by the reference signal.

7. The method according to claim 5 or 6, characterized in that a measurement or a calculation of exposure times (ti to tg) based on the last before or first after the respective image exposure synchronization pulse (I ₁ to I ₁₆ ) is obtained.

8. The method according to any one of claims 1 to 7, characterized in that for a camera (6) of the measuring device (4) a constant image exposure frequency is selected, which in a non-integral

Relative to the movement frequency of the object (2) stands.

9. The method according to claim 8, characterized in that during the detection of the subareas of the object surface by the camera (6) of the measuring device (4) respectively the time interval (ti to t ₉ ) between the detection time and the time at which the object ( 2) has reached, is measured and registered a reference position, where in which the contour data of the detected subregions are transformed into a common object coordinate system on the basis of the registered time intervals (ti to t ₉ ).

10. The method according to claim 9, characterized in that the time intervals (ti to tg) are measured by means of a real-time capable arithmetic unit.

11. The method according to claim 9 or 10, characterized in that for measuring the time intervals (ti to tg), the number of image clocks and the frame rate or the frame period of the camera (6) is used.

12. The method according to claim 8, characterized in that, when carrying out the measurement, the distance values between spatially adjacent subareas are continuously calculated and the measurement is ended as soon as the maximum value of all currently present distance values falls below a predetermined threshold value.

13. The method according to any one of claims 1 to 7, characterized in that for detecting the partial areas of the object surface at different instantaneous positions of the object (2) the camera (6) of the measuring device (4) each timed to a pre-calculated exposure time (ti to tg) is exposed.

14. The method according to claim 13, characterized in that in the prediction of the exposure times (ti to tg), the time interval between two successive exposure times (ti to tg) is selected to be larger than the image period (T) of a

Camera picture or camera picture frame.

15. The method according to claim 13 or 14, characterized in that for time-controlled exposure of the camera (6) to previously calculated exposure times (ti to tg) a real-time calculator is used.

16. The method according to any one of claims 13 to 15, characterized in that the exposure of the camera (6) by an external triggering of the camera (6) or an external triggering of a mechanical or electronic camera shutter is controlled.

17. The method according to any one of claims 13 to 15, characterized in that the exposure of the camera (6) on the switch-on and duty cycle of a light source (5) of the measuring device (4) is set.

18. The method according to claim 1, characterized in that for the process visualization a graphic is displayed which shows the position of the individual detected partial regions on a stylized representation of the object (2) or in one of the movement periods of the object (2). indicating scale (11), wherein preferably the graph is continuously updated during the measurement.